View Usage

After signing in to the console, open the Usage Detail page to view model usage details.

The Usage Detail page is best for inspecting individual requests, including token usage, cost, latency, and whether the request used streaming. To view aggregated metrics, usage trends, and latency performance over a selected time range, go to the Metrics page.

Filters

You can filter usage details by the following dimensions:

Model: View records for all models, or select a specific model.
API Key: View records generated by all keys, or select a specific API key.
Streaming: View all requests, or only streaming / non-streaming requests.
Time range: Select the last 24 hours, last 7 days, last 30 days, or define a custom range with a start date and end date.

After you adjust the filters, the page shows usage records for the selected time range and conditions. You can also choose Refresh to load the latest data.

Fields

The Usage Detail list shows model call details by request. Common fields include:

Field	Description
Model	The model called by the request.
API Key	The API key used to send the request.
Request Time	The time when the request occurred.
Tier	The model service tier used by the request.
Input	Input tokens that are not served from cache and need to be processed for the request, including message history, context material, tool definitions, and similar content. For vision and multimodal models, input tokens include text tokens and image tokens.
Image Input	Tokens converted from visual input. Image Input is part of input tokens.
Cache Read / Cache Write	Prompt Caching read and write token counts. Cache Read means reused cached input tokens; Cache Write means input tokens written into cache for later reuse.
Reasoning	Reasoning tokens, which are output tokens used by a reasoning model’s internal reasoning process.
Output	Output tokens, including reasoning tokens and tokens generated for the final visible answer.
Total Tokens	Total token usage for the request. It usually includes Input, Cache Read, Cache Write, and Output token counts.
Amount	Cost generated by the request.
E2E (ms)	Total time from request start to response completion.
Audio duration (s)	Audio duration for audio model requests.
Streaming	Whether the request used streaming output.
TTFT (ms)	Time from request start to the first output token.
Output Speed	Model output speed.
ms / token	Average time per output token.

Input, Cache Read, Cache Write, and Output can have separate unit prices. Check the model detail page in Models for exact pricing. In the current UI, reasoning tokens are already included in Output, so do not add Reasoning and Output together when analyzing cost. Reasoning tokens may or may not be returned in the API response, but they are included in platform usage statistics and billing. Image input is converted into input tokens and billed at the input unit price. For details, see Token Usage and Billing.

If there are no records under the current filters, the page shows that no records are available for the selected time range. You can expand the time range, or check whether the selected model, API key, or streaming filter is too narrow.

Get Started

Models

Usage

API Guides

Tool Integrations

Filters

Fields

​Filters

​Fields

Filters

Fields